Naci Yastikli

Influence of System Calibration on Direct Sensor Orientation

The determination of object coordinates based on direct sensor orientation is an extrapolation from the projection centers to the ground coordinate system. Like any extrapolation, it is sensitive to random and systematic errors, as well as, to improper data handling. Direct sensor orientation is based on the combination of an Inertial Measurement System (IMU) and GPS. The GPS antenna, the IMU, and the imaging sensor are located in different positions, and the latter two have different orientations. Therefore, the calibration of all sensors and the relation between the sensors is of vital importance for precise ground positioning. The system calibration includes the determination of the boresight misalignment, the interior camera orientation, and the GPS antenna offset. A rigorous mathematical model is required. The inner orientation of the camera used has to be determined under flight conditions to achieve sufficient results. In this paper, the influence of the system calibration to the direct sensor orientation and improper data handling will be shown.

Contribution of Normalized DSM to Automatic Building Extraction from HR Mono Optical Satellite Imagery

Abstract Building extraction from high resolution (HR) satellite imagery is one of the most significant issue for remote sensing community. Manual extraction process is onerous and time consuming thats why the improvement of the best automation is a crucial topic for the researchers. In this study, we aimed to expose the significant contribution of normalized digital surface model (nDSM) to the automatic building extraction from mono HR satellite imagery performing two-step application in an appropriate study area which includes various terrain formations. In first step, the buildings were manually and object-based automatically extracted from ortho-rectified pan-sharpened IKONOS and Quickbird HR imagery that have 1 m and 0.6 m ground sampling distances (GSD), respectively. Next, the nDSM was created using available aerial photos to represent the height of individual non-terrain objects and used as an additional channel for segmentation. All of the results were compared with the reference data, produced from aerial photos that have 5 cm GSD. With the contribution of nDSM, the number of extracted buildings was increased and more importantly, the number of falsely extracted buildings occurred by automatic extraction errors was sharply decreased, both are the main components of precision, completeness and overall quality.

Quantitative assessment of remotely sensed global surface models using various land classes produced from Landsat data in Istanbul

Digital elevation model (DEM) is the most popular product for three-dimensional (3D) digital representation of bare Earth surface and can be produced by many techniques with different characteristics and ground sampling distances (GSD). Space-borne optical and synthetic aperture radar (SAR) imaging are two of the most preferred and modern techniques for DEM generation. Using them, global DEMs that cover almost entire Earth are produced with low cost and time saving processing. In this study, we aimed to assess the Satellite pour l’observation de la Terre-5 (SPOT-5), High Resolution Stereoscopic (HRS), the Advanced Space-borne Thermal Emission and Reflection Radiometer (ASTER), and the Shuttle Radar Topography Mission (SRTM) C-band global DEMs, produced with space-borne optical and SAR imaging. For the assessment, a reference DEM derived from 1:1000 scaled digital photogrammetric maps was used. The study is performed in 100 km2 study area in Istanbul including various land classes such as open land, forest, built-up land, scrub and rough terrain obtained from Landsat data. The analyses were realized considering three vertical accuracy types as fundamental, supplemental, and consolidated, defined by national digital elevation program (NDEP) of USA. The results showed that, vertical accuracy of SRTM C-band DEM is better than optical models in all three accuracy types despite having the largest grid spacing. The result of SPOT-5 HRS DEM is very close by SRTM and superior in comparison with ASTER models.

ALS model-based comparison of Cosmo-SkyMed and TerraSAR-X HS DSMs on varied land forms

Since 2007, the revolutionary X-band synthetic aperture radar (SAR) satellites Cosmo-SkyMed (CSK) and TerraSAR-X (TSX) have been collecting high-resolution images that are convenient for digital surface model (DSM) acquisition by the interferometric SAR (InSAR) technique. In this study, the potential of DSMs derived from CSK and TSX high-resolution spotlight (HS) images is thoroughly analysed regarding the basic accuracy metrics absolute vertical accuracy (AVA) and relative vertical accuracy (RVA). Utilising convenient InSAR pairs, 2 m gridded DSMs are generated in Istanbul and validated with model-based comparisons using an actual airborne laser scanning (ALS) DSM. Results show that CSK and TSX DSMs are compatible in open, built-up and forest land forms. The AVAs are between 8 m and 10 m based on standard deviation of height discrepancies against the ALS model. The RVAs, calculated by the coherence of neighbouring pixels for each DSM, are superior to AVAs for both CSK and TSX.

Performance evaluation of thermographic cameras for photogrammetric documentation of historical buildings

Thermographic cameras record temperatures emitted by objects in the infrared region. These thermal images can be used for texture analysis and deformation caused by moisture and isolation problems. For accurate geometric survey of the deformations, the geometric calibration and performance evaluation of the thermographic camera should be conducted properly. In this study, an approach is proposed for the geometric calibration of the thermal cameras for the geometric survey of deformation caused by moisture. A 3D test object was designed and used for the geometric calibration and performance evaluation. The geometric calibration parameters, including focal length, position of principal point, and radial and tangential distortions, were determined for both the thermographic and the digital camera. The digital image rectification performance of the thermographic camera was tested for photogrammetric documentation of deformation caused by moisture. The obtained results from the thermographic camera were compared with the results from digital camera based on the experimental investigation performed on a study area.

Terrain estimation performance of advanced SAR satellites: Cosmo-SkyMed and TerraSAR-X

Presently, Cosmo-SkyMed (CSK) and TerraSAR-X (TSX) are the most advanced synthetic aperture radar (SAR) satellites which offer highest spatial resolution that could not be achieved from previous spaceborne SAR missions. In addition to high spatial resolution both of them are capable of repeat-pass interferometry that makes possible three dimensional digital modelling of visible earth surface from space. In this study, digital earth modelling performance of CSK and TSX high resolution spotlight (HS) images is comprehensively analyzed regarding absolute and relative vertical accuracies of generated digital surface models (DSMs). Accuracies were calculated on different land classes in Istanbul using a reference DSM derived by airborne laser scanning (ALS). Compatible results were achieved for CSK and TSX in open, built-up, and forest land classes. The absolute vertical accuracies are between 8 m and 10 m considering standard deviation. The relative vertical accuracies are superior to absolute both for CSK and TSX as expected from active sensing.

The Use of the Terrestrial Photogrammetry in Reverse Engineering Applications

The reverse engineering applications has gained great momentum in industrial production with developments in the fields of computer vision and computer-aided design CAD. The reproduction of an existing product or a spare part, reproduction of an existing surface, elimination of the defect or improvement of the available product are the goals of industrial reverse engineering applications. The first and the most important step in reverse engineering applications is the generation of the three dimensional 3D metric model of an existing product in computer environment. After this stage, many operations such as the preparation of molds for mass production, the performance testing, the comparison of the existing product with other products and prototypes which are available on the market are performed by using the generated 3D models. In reverse engineering applications, the laser scanner system or digital terrestrial photogrammetry methods, also called contactless method, are preferred for the generation of the 3D models. In particular, terrestrial photogrammetry has become a popular method since require only photographs for the 3-dimensional drawing, the generation of the dense point cloud using the image matching algorithms and the orthoimage generation as well as its low cost. In this paper, an industrial application of 3D information modelling is presented which concerns the measurement and 3D metric modelling of the ship model. The possible usage of terrestrial photogrammetry in reverse engineering application is investigated based on low cost photogrammetric system. The main aim was the generation of the dense point cloud and 3D line drawing of the ship model by using terrestrial photogrammetry, for the production of the ship in real size as a reverse engineering application. For this purpose, the images were recorded with digital SLR camera and orientations have been performed. Then 3D line drawing operations, point cloud and orthoimage generations have been accomplished by using PhotoModeler software. As a result of the proposed terrestrial photogrammetric steps, 0.5 mm spaced dense point cloud and orthoimage have been generated. The obtained results from experimental study were discussed and possible use of proposed methods was evaluated for reverse engineering application.